Evaluation of a Continuous Glucose Monitoring–Based Insulin Titration Protocol for Inpatients With Type 2 Diabetes in Nonintensive Care Unit Settings

Abstract

Background:

No widely adopted continuous glucose monitoring (CGM)-based insulin titration protocol exists, which may limit the effects of inpatient CGM on glycemic and clinical outcomes. We evaluate the acceptability and operability of the protocol proposed by Olsen et al for inpatients with type 2 diabetes in non–intensive care unit (non-ICU) settings.

Method:

7 inpatient diabetes team members, responsible for daily insulin titration, decided on insulin adjustments for 353 days. The members had the option to follow the CGM-based insulin protocol or override it for basal, prandial, and correctional insulin, separately, in 84 inpatients monitored by CGM. Questionnaires were used to evaluate the protocol’s operability by the teams.

Results:

Of 456 basal insulin titration decisions, 439 (96.3%) adhered to the protocol. For prandial insulin, adherence rates were 83.9% (125/149) for breakfast, 87.2% (130/149) for lunch, and 92.6% (138/149) for dinner (p=0.163). All correctional insulin titrations adhered to the protocol. All team members expressed a preference for having a protocol for CGM-based insulin titration and rated the protocol’s usability on a 1 to 10 scale, with mean scores (SD) of 8.7 (0.9) for basal insulin, 8.3 (1.4) for prandial insulin, and 7.4 (1.9) for correctional insulin.

Conclusions:

The CGM-based insulin titration protocol by Olsen et al has been successfully implemented for titrating basal, prandial, and correctional insulin in inpatients with type 2 diabetes in non-ICU settings. It was highly accepted by inpatient diabetes teams and provides a framework for effective CGM implementation in these settings.

Keywords

continuous glucose monitoring inpatient insulin titration hospital type 2 diabetes

Introduction

Effective management of in-hospital hyperglycemia relies on insulin therapy, yet creating individualized insulin regimens is challenging due to altered glucose metabolism during hospitalization and limited specialized diabetes care across wards.¹ The result is often high glycemic variability, hypoglycemia, and hyperglycemia, which are linked to increased morbidity, prolonged hospital stays, and higher mortality rates.^2-4

Inpatient standard-of-care glucose monitoring relies on point-of-care (POC) glucose testing by finger stick, typically performed 3 to 5 times daily, leaving significant data gaps, especially overnight. These gaps complicate insulin titration and increase the risk of dysglycemia and its associated complications.⁵ Continuous glucose monitoring (CGM) offers a promising alternative by providing real-time glucose data and a comprehensive glycemic profile. Despite its proven benefits in outpatient settings, CGM has shown limited success in hospitals likely due to lack of diabetes-trained staff, and lack of effective and operational CGM-based insulin titration protocols.^6-15

International guidelines advocate for protocolized CGM-based insulin titration, yet no standardized or widely adopted protocols currently exist.^13,16,17 This absence may lead to clinical inertia, where health care providers hesitate to make timely insulin adjustments even when real-time CGM data are available.¹⁸ Therefore, an operational CGM-based insulin titration protocol may be critical to realize CGM’s potential to improve inpatient glycemic control and clinical outcomes.^19,20

This analysis seeks to evaluate the acceptability and operability of the CGM-based insulin titration protocol by Olsen et al,¹⁸ as previously demonstrated in the DIAbetes TEam and Cgm (DIATEC) trial, which integrated protocolized CGM-based insulin titration with inpatient diabetes teams, resulting in significant clinical and glycemic improvements compared with POC glucose testing.²⁰ Furthermore, the protocol’s operability as perceived by inpatient diabetes teams is assessed.

Method

Data are from the DIATEC trial.²⁰ From April 11, 2023, to March 29, 2024, we enrolled 166 non-ICU medical and surgical patients with type 2 diabetes from Copenhagen University Hospital—North Zealand and Herlev-Gentofte Hospital, aged ≥18, and expected to stay in the hospital for at least 2 days. Key exclusions criteria included patients on long-acting basal insulin >24 hours (insulin glargine 300 U/mL or insulin degludec), those on systemic glucocorticoids (prednisone >5 mg/day), or with dialysis or an estimated glomerular filtration rate (eGFR) <15 mL/min/1.73 m². Additional criteria are published.¹⁸ In total, 84 patients were randomized to have glucose monitoring by CGM, and 82 patients randomized to have glucose levels assessed daily via POC glucose testing and insulin titration via a CGM- or POC-based insulin titration protocol, respectively.¹⁸

Each day, inpatient diabetes teams were required to decide on insulin titration for basal, prandial, and correctional insulin, using either the Olsen et al protocol (Table 1) or their clinical discretion. We assessed the adherence to the protocol by tracking all decisions related to titration of basal, prandial, and correctional insulin, separately, during hospitalization. We then determined whether these decisions followed the protocol or were modified by clinical discretion. If an insulin dose decision deviated from the protocol and was instead based on clinical discretion, we recorded the difference in insulin units between the recommended dose from the protocol and the actual dose administered. The inpatient team had the option to provide a comment when deviating from the protocol. Seven members of the inpatient diabetes team responsible for daily insulin titration completed questionnaires evaluating the operability of the protocol (Table 2).

Table 1.

CGM-Based Protocol by Olsen et al for Titrating Basal, Prandial, and Correctional Insulin.

Orders at admission
Discontinue the patient’s oral and non-oral glucose-lowering medicaments.
Insulin orders during admission
Basal insulin (mandatory order) • Basal insulin is given at lunch to allow same-day insulin titration decisions during the morning to take effect. Starting basal insulin dose is 0.25 U/kg/day, rounded to the nearest unit. Exceptions are: • Basal insulin dose is 0.20 U/kg/day in patients aged >75 years and/or with an eGFR ≤60 mL/min/1.73 m² and/or body mass index (BMI) ≤ 22.5 kg/m². • Basal insulin is continued on the usual out-hospital dose in patients treated with basal insulin at home. • In perioperative or fasting patients, basal insulin is paused on the day of surgery or during fasting and patients are managed by glucose-insulin-potassium (GIK) infusion. If the postoperative patient eats a full meal before 22:00 h, GIK infusion is stopped, and half basal insulin dose is given. The full basal insulin dose is continued the following days. Prandial insulin (ordered at the discretion of the inpatient diabetes teams when glucose levels >10.0 mmol/L [>180 mg/dL] are observed postprandially for 1-2 days) Starting total prandial insulin dose is 0.25 U/kg/day, rounded to the nearest unit, and given in 3 equally divided doses for breakfast, lunch, and dinner. Exceptions are: • Prandial insulin dose is 0.20 U/kg/day in patients aged >75 years and/or with an eGFR ≤60 mL/min/1.73 m² and/or body mass index (BMI) ≤22.5. • Prandial insulin is continued on the usual out-hospital dose in patients treated with prandial insulin at home. • In perioperative or fasting patients, prandial insulin is paused on the day of surgery or during fasting. Correctional insulin (mandatory order) Correctional insulin is given by the sliding scale insulin scheme below according to glucose levels at 03:00 h, preprandial at breakfast, lunch, and dinner, and before bedtime (22:00 h) or for patients not eating at 03:00 h., 08:00 h., 12:00 h., 17:00 h., and 22:00 h and/or hyperglycemic alarms for CGM-arm patients.
Glucose Level		Recommended insulinbreak/>(U)	Insulin change(U)
mmol/L	mg/dL	Recommended insulinbreak/>(U)	Insulin change(U)
<10.0	<180	0	0
10.0-11.9	180-214	4	± 2
12.0-15.9	215-286	6	± 3
16.0-19.9	287-358	8	± 4
≥ 20.0	> 358	10	± 6
Titrating basal, prandial, and correctional insulin
Titrating prandial insulin Titrated is done by the nadir of the between-meal CGM glucose levels. If there is no nadir on the CGM trace between meals, the pre-prandial glucose level guides the decision to titrate prandial insulin. • <3.0 [<54]: Decrease insulin dose by 30%. • 3.0-3.8 [54-69]: Decrease insulin dose by 20%. • 3.9-5.5 [70-99]: Decrease insulin dose by 10%. • 5.6-7.8 [100-140]: No changes. • 7.9-10.0 [141-180]: Increase insulin dose by 10% if no hypoglycemic events are observed the previous day. • 10.1-15.0 [181-270]: Increase insulin dose by 20% if no hypoglycemic events were observed the previous day. • >15.0 [>270]: Increase insulin dose by 30% if no hypoglycemic events were observed the previous day. Titrating basal insulin (mmol/L) [mg/dL] Titration of basal insulin (mmol/L) [mg/dL] is done by glucose level ranges from 00:00 h till breakfast (07:00h): • <3.0 mmol/L for >5% (>20 min) of glucose levels and/or at least 1 hypoglycemic event* <3.0 mmol/L: Decrease dose by 30%. • 3.0-3.8 mmol/L for >5% (>20 min) of glucose levels and/or at least 1 hypoglycemic event* 3.0-3.8 mmol/L: Decrease dose by 20%. • 3.9-5.5 mmol/L for >10% (>40 min) of glucose levels: Decrease dose by 10%. • 5.6-7.8 mmol/L: No change. • 7.9-10.0 mmol/L for >10% (>40 min) of glucose levels: Increase dose by 10%. • 10.1-15.0 mmol/L for >10% (>40 min) of glucose levels: Increase dose by 20%. • >15.0 mmol/L for >10% (>40 min) of glucose levels: Increase dose by 30%. Titrating correctional insulin Titration of sliding scale insulin is done based on the decision of whether the patient is insulin resistant (add insulin) or insulin sensitive (subtract insulin) by adjusting levels of Recommended insulin (U) by the column Insulin change (U). The trend in increasing or decreasing basal and/or prandial insulin might guide this decision. * A CGM-detected hypoglycemic event is defined as 3 consecutive CGM glucose levels in the hypoglycemic range.²¹

Table 2.

Questions for the Inpatient Diabetes Teams on the Operability and Efficacy of the Olsen et al¹⁸ CGM-Based Insulin Titration Protocol.

Questions regarding POC glucose testing vs. CGM for diabetes management
How easy do you find point-of-care glucose testing to be for adjusting insulin doses in noncritically ill inpatients with type 2 diabetes?	Very difficult Difficult Neutral Easy Very easy	n = 0 (0%) n = 0 (0%) n = 2 (29%) n = 4 (57%) n = 1 (14%)	.190
How easy do you find continuous glucose monitoring to be for adjusting insulin doses in noncritically ill inpatients with type 2 diabetes?	Very difficult Difficult Neutral Easy Very easy	n = 0 (0%) n = 0 (0%) n = 0 (0%) n = 5 (71%) n = 2 (29%)	.190
Do you feel confident using continuous glucose monitoring in managing glucose control for noncritically ill inpatients with type 2 diabetes?	Yes, n = 6 (86%)	No, n = 1 (14%)
Comment 1: “I am not confident using CGM because the staff are not critical of the pitfalls. For example, they do not assess whether the patient is dehydrated, and they do not check glucose levels. If the patient is self-sufficient and able to act, it’s fine. The staff forget to document glucose levels, as it does not automatically go into the patient’s electronic health record.”
Have you experienced any issues using continuous glucose monitoring for noncritically ill inpatients with type 2 diabetes?	Yes, n = 2 (29%)	No, n = 5 (71%)
Comment 1: ”No signal. For example, because the phone transmitting the signal was out of battery. Unreliable measurements due to pressure on the sensor. No signal in general due to poor internet connection.” Comment 2: ”Gaps in measurements—patients are out of range from the phone transmitting the signal. They sleep on their CGM [causing pressure artifacts]. The sensor needs to be removed for scanning. Staff do not check glucose levels when it’s under 5 or over 15 mmol/L.
What do you prefer to use for insulin titration for noncritically ill inpatients with type 2 diabetes?
In general	POC glucose testingCGMEqually effective	n = 0 (0%)n = 3 (43%)n = 4 (57%)	ref.
For basal insulin titration f	POC glucose testingCGMEqually effective	n = 0 (0%)n = 4 (57%)n = 3 (43%)	.143
For prandial insulin titration	POC glucose testingCGMEqually effective	n = 0 (0%)n = 6 (86%)n = 1 (14%)	.429
For correctional insulin titration	POC glucose testingCGMEqually effective	n = 0 (0%)n = 5 (71%)n = 2 (29%)	.430
Did you find the hypo- and hyperglycemia alarms helpful or bothersome in managing glucose levels for noncritically ill inpatients with type 2 diabetes?	CGM alarms were helpfulCGM alarms were bothersomeI did not receive any CGM alarms	n = 5 (71%)n = 1 (14%)n = 1 (14%)
Do you feel that continuous glucose monitoring data provided you with better insight into the patient’s glucose levels, leading to different clinical decisions regarding insulin titration and planning compared with point-of-care glucose testing?	Yes, n = 6 (86%)	No, n = 1 (14.3%)
Comment 1: “Generally felt more confident giving higher insulin doses, more “aggressive” treatment with CGM.” Comment 2: “Because of the ability to continuously monitor blood sugar, insulin could be adjusted more quickly according to the algorithm (with peace of mind), as fluctuations would show up promptly.” Comment 3: “Daily monitoring and decision-making regarding titration meant that patients often reached their blood sugar targets more quickly.” Comment 4: “Adjustment from day to day, either by altering the insulin dose slightly or by not taking action unnecessarily.” Comment 5: “Gained a completely new insight into the daily glucose profile—when they dipped, how they were in the morning. Sometimes, staff measured fasting blood sugar too late. Also felt more confident giving more insulin, knowing the alarm would go off.”
Questions regarding the CGM-based insulin titration algorithm
Indicate whether the following statement is true or false: I felt more confident in initiating prandial insulin for patients on continuous glucose monitoring rather than point-of-care glucose testing, as I knew hypoglycemia would be detected either the same day through alarms or by me or a colleague the following day during insulin titration.	Yes, n = 7 (100%)	No, n = 0 (0%)
Do you prefer an algorithm to support your existing clinical knowledge for glucose management and insulin titration for inpatients with type 2 diabetes who have their glucose levels measured by continuous glucose monitoring?	Yes, n = 7 (100%)	No, n = 0 (0%)
On a scale from 1 to 10, how helpful has the continuous glucose monitoring-based insulin titration protocol been in adjusting insulin for noncritically ill inpatients with type 2 diabetes? mean (SD)
In general	8.7 (0.9)
For basal insulin	8.7 (0.9)
For prandial insulin	8.3 (1.4)
For correctional insulin	7.4 (1.9)
Additional comments from the inpatient diabetes team on the CGM-based insulin titration protocol
Comment 1: “A prerequisite for the algorithm is the calculation of the percentage of time spent in hypo-, normo-, and hyperglycemia. The percentages have probably helped me more than the curves themselves. AI (artificial intelligence, red.) should be able to help suggest insulin treatments based on CGM curves.” Comment 2: “Good idea to use CGM during hospitalization. It would likely help prevent hypo- and hyperglycemia on a large scale. NOTE: false alarms can occur when the patient is cold (skin is cold) or lying on the CGM.” Comment 3: “It has been very useful to receive suggestions for up- and down-titration of insulin. However, in several cases, I have also used my clinical experience and not always followed the algorithm, but it’s certainly a helpful tool.” Comment 4: “I felt more confident in tightening insulin management for patients with CGM, as we could see data from the entire day and could confirm there was no hypoglycemia.”

Abbreviations: CGM, continuous glucose monitoring; POC, point-of-care; SD, standard deviation.

Outcomes

The primary outcome in this analysis was to assess the degree of adherence to the Olsen et al CGM-based insulin titration protocol. Secondary outcomes included a questionnaire-based evaluation (Table 2) of the protocol by the inpatient diabetes teams applying the protocol.

Description of the CGM-Based Insulin Titration Protocol

At admission, noninsulin antidiabetic medications were discontinued, and patients were initiated on a basal-bolus insulin regimen of 0.4 to 0.5 U/kg/day distributed equally between basal insulin and prandial insulin for the 3 main meals.^18,22 Prandial insulin was ordered at the discretion of the inpatient diabetes teams when glucose levels >10.0 mmol/L were observed postprandially for 1 to 2 days. Correctional insulin was also administered.

In general, daily insulin adjustments adhered to the recommendations from the Olsen et al protocol, though the recommendations could be overruled at the clinical discretion of the inpatient diabetes teams when deemed necessary. Basal insulin was increased for sustained fasting (nighttime) hyperglycemia, and prandial or correctional insulin was initiated or increased for hyperglycemia during the day. CGM-detected hypo- (<3.9 mmol/L) and hyperglycemia (>13.9 mmol/L for 2 hours) alarms were activated and notified the inpatient diabetes teams. POC glucose testing was used to verify hypoglycemia due to reduced CGM sensor accuracy at lower glucose levels.²³ CGM trend arrows were not considered. The full CGM-based insulin titration protocol by Olsen et al is outlined in Table 1.

Rules for CGM-Based Insulin Titration

Two rules applied to CGM-based insulin titration in a prioritized order.

Rule of the lowest

Prioritizes low glucose levels when both low and high glucose levels occur during a titration period, provided both exceed the percentage-based thresholds in Table 1. This rule directs a reduction in basal and/or prandial insulin to prevent hypoglycemia.

Rule of the extremes

Extreme glucose levels that deviate from the titration range of 5.6 to 7.8 mmol/L (100-140 mg/dL)—whether consistently below, consistently above, or partially within and outside the titration range—are prioritized, for example, level 2 (>13.9 mmol/L) over level 1 (10.0-13.9 mmol/L) hyperglycemia, as the first range represents a more significant (“extreme”) deviation from the target range.

Ethical Statement

Before enrollment, all subjects provided informed written consent. The Ethics Committee of the Capital Region of Denmark approved the study (journal number 2301240). The DIATEC trial was prospectively registered with ClinicalTrials.gov as NCT05803473.

Statistical Analyses

Categorical variables were compared with the chi-squared test and presented as n (%). A generalized estimation equation model was applied to test whether there was a development over time in whether the CGM-based insulin titration protocol was followed or overruled by clinical decisions. As the inpatient diabetes teams were off shifts during weekends, and no insulin titration decisions were made, only insulin titration decisions from weekdays were analyzed. Linear mixed regression models were used to model the impact of nonadherence to basal and/or bolus (prandial) insulin titration on changes in consensus CGM-outcomes¹³ compared with within-subject glycemic periods when insulin titration adhered to the protocol. Consensus CGM-outcomes included time in range (TIR), 3.9 to 10.0 mmol/L; 3 levels of time above range (TAR) 10.0 to 13.9 mmol/L, >13.9 mmol/L, and >10.0 mmol/L; 3 levels of time below range (TBR) 3.0 to 3.9 mmol/L, <3.0 mmol/L, and <3.9 mmol/L; mean glucose level; standard deviation (SD) of the CGM glucose distribution, and coefficient of variation (CV. hypoglycemic events were not modeled due to zero-inflation of the outcome. For basal insulin titration nonadherence, glycemic status from the following night was paired with the preceding nonadherence event, and periods of protocol adherence were also paired accordingly. Similarly, for bolus (prandial) insulin nonadherence, glycemic status from the subsequent 4-hour period was paired with the nonadherence event, with protocol-adherent glycemic periods following the same structure. The linear mixed regression analyses were adjusted for the day of admission. Statistical significance was set at a 2-sided P-value ≤.05.

Results

Baseline of Inpatients With Type 2 Diabetes

Baseline characteristics are published elsewhere.²⁰ Briefly, 84 subjects were managed by CGM during hospitalization with a mean (SD) age of 76.6 (9.5) years, of which 52 (61.9%) were males. The duration of diabetes was on average (SD) 13.0 (8.6) years and 17 (20.2%) were treated with insulin therapy before hospitalization.

Baseline Characteristics of the Inpatient Diabetes Teams

Two individuals (29%) were from the Copenhagen University Hospital—North Zealand, and 5 individuals (71%) were from Herlev-Gentofte Hospital with an average (SD) experience of 16.0 (4.9) years as working with inpatient diabetes management being employed at the current workplace of 10.9 years (4.2) and with 8.0 (4.6) years as an inpatient diabetes specialist.

Evaluation of CGM-Based Insulin Titration Protocol

Among 456 basal insulin titration decisions made for 84 subjects during hospitalization, 439 (96.3%) adhered to the protocol. The decisions that deviated from the protocol regarding basal insulin were, on median (IQR), 2.3 (4.0) U lower than the amount proposed by the protocol (Table 1) and the mean (SD) dose of basal insulin was 17.6 (9.7) U/day. Out of 149 prandial insulin decisions made for breakfast, lunch, and dinner in 41 subjects, the protocol was followed in 125 (83.9%), 130 (87.2%), and 138 (92.6%) instances, respectively (P = .163). The decisions that deviated from the protocol regarding breakfast, lunch, and dinner prandial insulin, respectively, were median (IQR) 1.0 (1.0) U, 1.0 (1.0), and 2.0 (1.0) lower than the amount proposed by the Olsen et al protocol and the mean (SD) dose of basal insulin was 5.8 (3.7) U/day. All correctional insulin doses adhered to the protocol. All deviations from the protocol regarding prandial and basal insulin were due to fear of hypoglycemia as expressed by the inpatient diabetes teams.

There was no development over time in whether the Olsen et al protocol was followed or overruled by clinical decisions regarding basal insulin (P = .954), breakfast prandial insulin (P = .941), lunch prandial insulin (P = .702), and dinner prandial insulin (P = .338). The same pattern was observed with no statistically significant differences across both sites and between surgical and medical wards (data not shown). No significant differences were observed in outcomes, including TIR, TAR, TBR, mean glucose, SD, or CV, when comparing glycemic periods with protocol adherence to those with nonadherence (all P > .05, data not shown).

CGM alarms for hypoglycemic events (<3.9 mmol/L) occurring prior to the usual time stamps for POC glucose testing with corrective insulin were triggered in 12 patients in the CGM arm (14.3%). Among these, 5 patients had 1 alarm, 4 had 2 alarms, 2 had 4 alarms, and 1 had 8 alarms. For hyperglycemic events (>13.9 mmol/L lasting at least 2 hours), alarms were triggered in 22 patients (26.2%), with 12 patients having 1 alarm, 4 having 2 alarms, 3 having 3 alarms, 1 having 4 alarms, and 2 having 6 alarms.

Experience From the Inpatient Diabetes Teams With the CGM-Based Insulin Titration Protocol

The results are summarized in Table 2. All 7 (100%) inpatient diabetes team members preferred having a protocol for CGM-based insulin titration. They rated its usability for titrating basal, prandial, and correctional insulin on a 1 to 10 scale on average (SD) as 8.7 (0.9), 8.3 (1.4), and 7.4 (1.9), respectively. The inpatient diabetes team found the Olsen et al protocol as easy to use as the POC glucose testing-based insulin titration protocol (P = .190).

Six inpatient diabetes team members (86%) expressed confidence in using CGM for managing glucose control in inpatients with type 2 diabetes, although signal loss in the telemetric CGM-setup²⁴ was mentioned in the commentaries as a primary issue with inpatient CGM use (eg, due to Bluetooth disconnections between the CGM and the smartphone running the Dexcom G6 app, internet connectivity issues between the smartphone and the Dexcom Clarity webpage, sensor or transmitter malfunctions, etc.) Preferences for CGM over POC glucose testing were observed in 4 (57%) inpatient diabetes team members for basal insulin, 6 (86%) for prandial, and 5 (71%) for correctional insulin titration. The remaining inpatient diabetes team members found CGM and POC equally effective in titrating basal, prandial, and correctional insulin, with no one favoring POC over CGM.

Five inpatient diabetes team members (71%) found hypo- and hyperglycemia alarms helpful in managing glucose levels, and 6 (86%) agreed that CGM data provided better insights into glucose patterns, influencing insulin titration and clinical planning. All inpatient diabetes team members reported feeling more confident initiating prandial insulin for patients monitored with CGM than POC glucose testing.

Discussion

In this study, we aimed to evaluate the Olsen et al CGM-based insulin titration protocol for inpatients with type 2 diabetes in non-ICU settings. Adherence to the protocol (Table 1) was consistently high (around 90%), reflecting strong alignment between the recommendations of the CGM-based protocol and the clinical judgment of inpatient diabetes teams regarding insulin titration. The protocol was well-received, deemed easy to implement, and preferred by inpatient diabetes teams who appreciated having a structured approach for CGM-based insulin titration.

There was no significant change over time in the degree to which the protocol was followed or overridden by clinical decisions for basal, prandial, or correctional insulins. This consistency indicates the protocol’s robustness and reliability, as it remained a trusted tool throughout the testing period of 353 days, even as the inpatient diabetes teams gained greater familiarity and expertise with CGM-based insulin titration and glucose management. No significant differences in glycemic outcomes were observed when comparing periods of protocol adherence to nonadherence, suggesting that minor deviations from the protocol (approximately 2 U/day for basal insulin and 1-2 U per meal for bolus/prandial insulin) did not result in harm. However, a CGM-based protocol for inpatient insulin titration may still be essential for optimizing the overall effectiveness of inpatient CGM.¹⁹

The DIATEC trial is currently the only trial showing clinically significant glycemic and clinical benefits of CGM over POC in inpatient non-ICU settings, which may be attributed mainly to the CGM-based insulin titration protocol by Olsen et al²⁰ The findings from the present analysis highlight the critical importance of an operational CGM-based insulin titration protocol if CGM in an inpatient setting is to have clinical and glycemic benefits compared with standard-of-care POC glucose testing.

In other published CGM-based insulin titration protocols in non-ICU settings,^14,24-30 the Olsen et al¹⁸ protocol stands out as being the only protocol making rules for how to act when glucose levels were both below, within, and above the target range in a single titration period, such as overnight for basal insulin titration by defining the “rule of lowest” and “rule of the extreme.” This may significantly guide decision-making, as healthcare providers otherwise may be unsure whether to decrease, maintain, or increase insulin, respectively.

The Olsen et al protocol also has a potential advantage in incorporating percentage-based thresholds for time spent within specific glycemic ranges before insulin titration is adjusted. This contrasts with the traditional 1-dimensional approach in other protocols,^14,24,29-31 which rely solely on glucose level cutoffs (eg, “if fasting glucose is 10.1-15.0 mmol/L, increase basal insulin by 20%”). Instead, the Olsen et al protocol adopts a 2-dimensional approach that considers both glucose levels and the percentage of time spent within these ranges (eg, “if fasting glucose is 10.1-15.0 mmol/L for ≥10% of readings, increase basal insulin by 20%”). The latter approach most likely enables more standardized insulin titration by viewing CGM data as time series data and not dichotomous data.

Strengths, Limitations, and Future Directions

We evaluated a CGM-based insulin titration protocol in a 2-center randomized controlled trial setup in different inpatient medical and orthopaedic settings, increasing the external validity of our findings.

Our study has limitations, including a lack of recruitment of patients receiving nutritional therapies or glucocorticoids during hospitalization, all populations which might be at great risk of developing dysglycemia during hospitalization³² and, therefore, could have altered the adherence to a fixed CGM-based insulin titration algorithm. Additionally, the mean (SD) age of participants, 76.1 (9.8) years, may limit the generalizability of these findings to younger hospitalized patients with diabetes. Also, the acceptability of the proposed CGM-based insulin titration protocol should be evaluated in other settings and by more inpatient diabetes teams.

Conclusion

The Olsen et al¹⁸ CGM-based insulin titration protocol has been successfully evaluated for titration of basal, prandial, and correctional insulin for inpatients with type 2 diabetes in non-ICU settings. The protocol was highly accepted by inpatient diabetes teams and offers a framework for effectively implementing CGMs in non-ICU inpatient settings, leading to significant clinical and glycemic improvements in outcomes compared with POC glucose testing.

Footnotes

Abbreviations

BMI, body mass index; CGM, continuous glucose monitoring; DIATEC, DIAbetes TEam and Cgm; eGFR, estimated glomerular filtration rate; GIK, glucose-insulin-potassium; ICU, intensive care unit; IQR, inter quartile range; POC, point-of-care; SD, standard deviation.

Author Contributions

MTO conceptualized, designed, and conducted the DIATEC trial with assistance from UPB, CKK, BL, JAA, HG, SL, KBH, and PLK. SHJ and LMT assisted in data collection. MTO drafted the first version of the manuscript. All authors read and approved the final manuscript.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: JAA: Has received speakers fees from Juzo, Mediq, KCI, and AstraZeneca. UPB: Has received speakers fees from Sanofi and Novo Nordisk and has served on advisory boards for Sanofi, Novo Nordisk, and Vertex. Additional authors declare no conflict of interest.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Novo Nordisk Foundation (grant no. NNF20SA0062872), Vissing Fonden, Jascha Fonden, Copenhagen University Hospital North Zealand Research Grant, Fru Olga Bryde Nielsen Fonden, PhD-START- puljen North Zealand Hospital, A.P. Møller Fonden til Lægevidenskabens Fremme, A & J C Tvergaards Fond, Læge Inger Goldmanns Fond, Ellen Margrethe Cramers Fond, and Torben og Alice Frimodts Fond.

ORCID iDs

Mikkel Thor Olsen

Carina Kirstine Klarskov

Availability of Data and Materials

The data sets generated and/or analyzed during the DIATEC trial are available from the corresponding author upon reasonable request.

References

NaDIA Team. NHS national diabetes inpatient audit (NaDIA) – 2017 [Internet]. Published 2020. Accessed June 21, 2024. https://digital.nhs.uk/data-and-information/publications/statistical/national-diabetes-inpatient-audit/national-diabetes-inpatient-audit-nadia-2017

Umpierrez

Isaacs

Bazargan

You

Thaler

Kitabchi

. Hyperglycemia: an independent marker of in-hospital mortality in patients with undiagnosed diabetes. J Clin Endocrinol Metab. 2002;87(3):978-982.

Olsen

Dungu

Klarskov

Jensen

Lindegaard

Kristensen

. Glycemic variability assessed by continuous glucose monitoring in hospitalized patients with community-acquired pneumonia. BMC Pulm Med. 2022;22(1):83.

Baker

Janaway

Philips

, et al. Hyperglycaemia is associated with poor outcomes in patients admitted to hospital with acute exacerbations of chronic obstructive pulmonary disease. Thorax. 2006;61(4):284-289.

Rajendran

Kerry

Rayman

. Temporal patterns of hypoglycaemia and burden of sulfonylurea-related hypoglycaemia in UK hospitals: a retrospective multicentre audit of hospitalised patients with diabetes. BMJ Open. 2014;4(7):1-8.

Umpierrez

Klonoff

. Diabetes technology update: use of insulin pumps and continuous glucose monitoring in the hospital. Diabetes Care. 2018;41(8):1579-1589.

Seisa

Saadi

Nayfeh

, et al. A systematic review supporting the endocrine society clinical practice guideline for the management of hyperglycemia in adults hospitalized for noncritical illness or undergoing elective surgical procedures. J Clin Endocrinol Metab. 2022;107(8):2139-2147.

Akiboye

Sihre

Al Mulhem

Rayman

Nirantharakumar

Adderley

. Impact of diabetes specialist nurses on inpatient care: a systematic review. Diabet Med. 2021;38(9):e14573.

Klarskov

Kristensen

. Experience from implementing telemetric in-hospital continuous glucose monitoring during the COVID-19 pandemic. J Diabetes Sci Technol. 2021;15(3):715-716.

10.

Wallia

Umpierrez

Rushakoff

, et al. Consensus statement on inpatient use of continuous glucose monitoring. J Diabetes Sci Technol. 2017;11(5):1036-1044.

11.

Peters

Ahmann

Battelino

, et al. Diabetes technology—continuous subcutaneous insulin infusion therapy and continuous glucose monitoring in adults: an endocrine society clinical practice guideline. J Clin Endocrinol Metab. 2016;101(11):3922-3937.

12.

Fonseca

Grunberger

Anhalt

, et al. Continuous glucose monitoring: a consensus Conference of the American Association of Clinical Endocrinologists and American College of Endocrinology. Endocr Pract. 2016;22(8):1008-1021.

13.

Spanakis

Cook

Kulasa

, et al. A consensus statement for continuous glucose monitoring metrics for inpatient clinical trials. J Diabetes Sci Technol. 2023;17(6):1527-1552.

14.

Spanakis

Urrutia

Galindo

, et al. Continuous glucose monitoring—guided insulin administration in hospitalized patients with diabetes: a randomized clinical trial. Diabetes Care. 2022;45(10):2369-2375.

15.

Cavalcante Lima Chagas

Teixeira

Clemente

MRC

, et al. Use of continuous glucose monitoring and point-of-care glucose testing in hospitalized patients with diabetes mellitus in non-intensive care unit settings: a systematic review and meta-analysis of randomized controlled trials. Diabetes Res Clin Pract. 2025;220:111986.

16.

Galindo

Umpierrez

Rushakoff

, et al. Continuous glucose monitors and automated insulin dosing systems in the hospital consensus guideline. J Diabetes Sci Technol. 2020;14(6):1035-1064.

17.

Shaw

JLV

Bannuru

Beach

, et al. Consensus considerations and good practice points for use of continuous glucose monitoring systems in hospital settings. Diabetes Care. 2024;47(12):2062-2075.

18.

Olsen

Klarskov

Bjergaard

Hansen

Kristensen

. Summary of clinical investigation plan for The DIATEC trial: inhospital diabetes management by a diabetes team and continuous glucose monitoring or point of care glucose testing—a randomised controlled trial. BMC Endocr Disord. 2024;24(1):60.

19.

Olsen

Liarakos

Wilmot

, et al. Implementation strategies for inpatient continuous glucose monitoring-based diabetes management: a systematic review [Ahead of print]. J Clin Endocrinol Metab. doi:10.1210/clinem/dgaf074.

20.

Olsen

Klarskov

Jensen

, et al. In-hospital DIAbetes management by a diabetes TEam and insulin titration algorithms based on Continuous glucose monitoring or point-of-care glucose testing in patients with type 2 diabetes (DIATEC)—a randomised controlled trial. Diabetes Care. 2025;48(4):1-11.

21.

Battelino

Alexander

Amiel

, et al. Continuous glucose monitoring and metrics for clinical trials: an international consensus statement. Lancet Diabetes Endocrinol. 2022;11(1):42-57.

22.

Umpierrez

Smiley

Zisman

, et al. Randomized study of basal-bolus insulin therapy in the inpatient management of patients with type 2 diabetes (RABBIT 2 Trial). Diabetes Care. 2007;30(9):2181-2186.

23.

Buschur

Faulds

Dungan

. CGM in the hospital: is it ready for prime time? Curr Diab Rep. 2022;22(9):451-460.

24.

Singh

Satyarengga

Marcano

, et al. Reducing inpatient hypoglycemia in the general wards using real-time continuous glucose monitoring: the glucose telemetry system, a randomized clinical trial. Diabetes Care. 2020;43(11):2736-2743.

25.

Hirsch

Draznin

Buse

, et al. Results from a randomized trial of intensive glucose management using CGM versus usual care in hospitalized adults with type 2 diabetes: the TIGHT study. Diabetes Care. 2025;48(1):118-124.

26.

Thabit

Rubio

Karuppan

, et al. Use of real-time continuous glucose monitoring in non-critical care insulin-treated inpatients under non-diabetes speciality teams in hospital: a pilot randomized controlled study. Diabetes Obes Metab. 2024;26(11):5483-5487.

27.

Dillmann

Amoura

Fall Mostaine

Coste

Bounyar

Kessler

. Feasibility of real-time continuous glucose monitoring telemetry system in an inpatient diabetes unit: a pilot study. J Diabetes Sci Technol. 2022;16(4):955-961.

28.

Klarskov

Windum

Olsen

, et al. Telemetric continuous glucose monitoring during the COVID-19 pandemic in isolated hospitalized patients in Denmark: a randomized controlled exploratory trial. Diabetes Technol Ther. 2022;24(2):102-112.

29.

Fortmann

Spierling Bagsic

Talavera

, et al. Glucose as the fifth vital sign: a randomized controlled trial of continuous glucose monitoring in a non-ICU hospital setting. Diabetes Care. 2020;43(11):2873-2877.

30.

Spanakis

Levitt

Siddiqui

, et al. The effect of continuous glucose monitoring in preventing inpatient hypoglycemia in general wards: the glucose telemetry system. J Diabetes Sci Technol. 2018;12(1):20-25.

31.

Singh

Levitt

Satyarengga

, et al. Continuous glucose monitoring in general wards for prevention of hypoglycemia: results from the glucose telemetry system pilot study. J Diabetes Sci Technol. 2020;14(4):783-790.

32.

Olsen

Klarskov

Hansen

Pedersen-Bjergaard

Kristensen

. Risk factors at admission of in-hospital dysglycemia, mortality, and readmissions in patients with type 2 diabetes and pneumonia. J Diabetes Complications. 2024;38(8):108803.